Main part of the disk accounting rewrite.

This is a wholesale rewrite of the existing disk space accounting, which
relies on percepu counters that are sharded by journal buffer, and
rolled up and added to each journal write.

With the new scheme, every set of counters is a distinct key in the
accounting btree; this fixes scaling limitations of the old scheme,
where counters took up space in each journal entry and required multiple
percpu counters.

Now, in memory accounting requires a single set of percpu counters - not
multiple for each in flight journal buffer - and in the future we'll
probably also have counters that don't use in memory percpu counters,
they're not strictly required.

An accounting update is now a normal btree update, using the btree write
buffer path. At transaction commit time, we apply accounting updates to
the in memory counters, which are percpu counters indexed in an
eytzinger tree by the accounting key.
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>

Teach the btree write buffer how to accumulate accounting keys - instead
of having the newer key overwrite the older key as we do with other
updates, we need to add them together.

Also, add a flag so that write buffer flush knows when journal replay is
finished flushing accounting, and teach it to hold accounting keys until
that flag is set.
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>

Until accounting keys hit the btree, they are deltas, not new versions
of the existing key; this means we have to teach journal replay to
accumulate them.

Additionally, the journal doesn't track precisely which entries have
been flushed to the btree; it only tracks a range of entries that may
possibly still need to be flushed.

That means we need to compare accounting keys against the version in the
btree and only flush updates that are newer.

There's another wrinkle with the write buffer: if the write buffer
starts flushing accounting keys before journal replay has finished
flushing accounting keys, journal replay will see the version number
from the new updates and updates from the journal will be lost.

To avoid this, journal replay has to flush accounting keys first, and
we'll be adding a flag so that write buffer flush knows to hold
accounting keys until then.
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>

New key type for the disk space accounting rewrite.

 - Holds a variable sized array of u64s (may be more than one for
   accounting e.g. compressed and uncompressed size, or buckets and
   sectors for a given data type)

 - Updates are deltas, not new versions of the key: this means updates
   to accounting can happen via the btree write buffer, which we'll be
   teaching to accumulate deltas.
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>

This updates bcachefs to use the new mount API:

- Update the file_system_type to use the new init_fs_context()
  function.

- Define the new fs_context_operations functions.

- No longer register bch2_mount() and bch2_remount(); these are now
  called via the new fs_context functions.

- Define a new helper type, bch2_opts_parse that includes a struct
  bch_opts and additionally a printbuf used to save options that can't
  be parsed until after the FS is opened. This enables us to parse as
  many options as possible prior to opening the filesystem while saving
  those options that need the open FS for later parsing.
Signed-off-by: Thomas Bertschinger <tahbertschinger@gmail.com>
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>

This introduces a new error code, option_needs_open_fs, which is used to
indicate that an attempt was made to parse a mount option prior to
opening a filesystem, when that mount option requires an open filesystem
in order to be validated.

Returning this error results in bch2_parse_one_mount_opt() saving that
option for later parsing, after the filesystem is opened.
Signed-off-by: Thomas Bertschinger <tahbertschinger@gmail.com>
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>

Mount options that take the name of a device that may be part of a
filesystem, for example "metadata_target", cannot be validated until
after the filesystem has been opened. However, an attempt to parse those
options may be made prior to the filesystem being opened.

This change adds a printbuf parameter to bch2_parse_mount_opts() which
will be used to save those mount options, when they are supplied prior
to the FS being opened, so that they can be parsed later.

This functionality is not currently needed, but will be used after
bcachefs starts using the new mount API to parse mount options. This is
because using the new mount API, we will process mount options prior to
opening the FS, but the new API doesn't provide a convenient way to
"replay" mount option parsing. So we save these options ourselves to
accomplish this.

This change also splits out the code to parse a single option into
bch2_parse_one_mount_opt(), which will be useful when using the new
mount API which deals with a single mount option at a time.
Signed-off-by: Thomas Bertschinger <tahbertschinger@gmail.com>
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>

New on disk format version for bch_alloc->stripe_sectors and
BCH_DATA_unstriped - accounting for unstriped data in stripe buckets.

Upgrade/downgrade requires regenerating alloc info - but only if erasure
coding is in use.
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>

Add a new pseudo data type, to track buckets that are members of a
stripe, but have unstriped data in them.
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>

Add a separate counter to bch_alloc_v4 for amount of striped data; this
lets us separately track striped and unstriped data in a bucket, which
lets us see when erasure coding has failed to update extents with stripe
pointers, and also find buckets to continue updating if we crash mid way
through creating a new stripe.
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>

Consolidate duplicated checks for extents/dirents/xattrs - these keys
should all have a corresponding inode of the correct type.
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>

The output of mount options such as "metadata_target" in `/proc/mounts`
uses the full path to the device.

mount(8) from util-linux uses the output from `/proc/mounts` to pass
existing mount options when performing a remount, so bcachefs should
accept as input the same form that it prints as output.

Without this change:

$ mount -t bcachefs -o metadata_target=vdb /dev/vdb /mnt
$ strace mount -o remount /mnt
...
fsconfig(4, FSCONFIG_SET_STRING, "metadata_target", "/dev/vdb", 0) = -1 EINVAL (Invalid argument)
...
Signed-off-by: Thomas Bertschinger <tahbertschinger@gmail.com>
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>

Add a new helper to fix inode_to_text()
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>

When "read_only" is exposed as a mount option, it is redundant with the
standard option "ro" and gives users multiple ways to specify that a
bcachefs filesystem should be mounted read-only. This presents the risk
of having inconsistent options specified.

This can be seen when remounting a read-only filesystem in read-write
mode, using mount(8) from util-linux. Because mount(8) parses the
existing mount options from `/proc/mounts` and applies them when
remounting, it can end up applying both "read_only" and "rw":

$ mount img -o ro /mnt
$ strace mount -o remount,rw /mnt
...
fsconfig(4, FSCONFIG_SET_FLAG, "read_only", NULL, 0) = 0
fsconfig(4, FSCONFIG_SET_FLAG, "rw", NULL, 0) = 0
...

Making "read_only" no longer a mount option means this edge case cannot
occur.

Fixes: 62719cf3 ("bcachefs: Fix nochanges/read_only interaction")
Signed-off-by: Thomas Bertschinger <tahbertschinger@gmail.com>
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>

A subsequent change will remove "read_only" as a mount option in favor
of the standard option "ro", meaning the userspace fsck command cannot
pass it to the fsck ioctl. Instead, in offline fsck, set "read_only"
kernel-side without trying to parse it as a mount option.

For compatibility with versions of the "bcachefs fsck" command that try
to pass the "read_only" mount opt, remove it from the mount options
string prior to parsing when it is present.
Signed-off-by: Thomas Bertschinger <tahbertschinger@gmail.com>
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>

this is for the userspace metadata dump tool
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>

Use try_cmpxchg() family of functions instead of
cmpxchg (*ptr, old, new) == old. x86 CMPXCHG instruction returns
success in ZF flag, so this change saves a compare after cmpxchg
(and related move instruction in front of cmpxchg).

Also, try_cmpxchg() implicitly assigns old *ptr value to "old" when
cmpxchg fails. There is no need to re-read the value in the loop.

No functional change intended.
Signed-off-by: Uros Bizjak <ubizjak@gmail.com>
Cc: Kent Overstreet <kent.overstreet@linux.dev>
Cc: Brian Foster <bfoster@redhat.com>
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>

fs-common.h needs dirent.h for enum bch_rename_mode
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>

Using filemap_read() can reduce unnecessary code execution
for non IOCB_DIRECT paths.
Signed-off-by: Youling Tang <tangyouling@kylinos.cn>
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>

Before patch:
```
 #cat btrees/inodes/keys
 u64s 17 type inode_v3 0:4096:U32_MAX len 0 ver 0:   mode=40755
   flags= (16300000)
   bi_size=0
```

After patch:
```
 #cat btrees/inodes/keys
 u64s 17 type inode_v3 0:4096:U32_MAX len 0 ver 0:
   mode=40755
   flags=(16300000)
   bi_size=0
```
Signed-off-by: Youling Tang <tangyouling@kylinos.cn>
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>

Fixed missing spaces displayed in journal_entry_dev_usage_to_text
while adjusting the display format to improve readability.

before:
```
 # bcachefs list_journal -a -t alloc:1:0 /dev/sdb
 ...
     dev_usage: dev=0free: buckets=233180 sectors=0 fragmented=0sb: buckets=13 sectors=6152 fragmented=504journal: buckets=1847 sectors=945664 fragmented=0btree: buckets=20 sectors=10240 fragmented=0user: buckets=1419 sectors=726513 fragmented=15cached: buckets=0 sectors=0 fragmented=0parity: buckets=0 sectors=0 fragmented=0stripe: buckets=0 sectors=0 fragmented=0need_gc_gens: buckets=0 sectors=0 fragmented=0need_discard: buckets=1 sectors=0 fragmented=0
```

after:
```
 # bcachefs list_journal -a -t alloc:1:0 /dev/sdb
 ...
     dev_usage: dev=0
       free: buckets=233180 sectors=0 fragmented=0
       sb: buckets=13 sectors=6152 fragmented=504
       journal: buckets=1847 sectors=945664 fragmented=0
       btree: buckets=20 sectors=10240 fragmented=0
       user: buckets=1419 sectors=726513 fragmented=15
       cached: buckets=0 sectors=0 fragmented=0
       parity: buckets=0 sectors=0 fragmented=0
       stripe: buckets=0 sectors=0 fragmented=0
       need_gc_gens: buckets=0 sectors=0 fragmented=0
       need_discard: buckets=1 sectors=0 fragmented=0
```
Signed-off-by: Youling Tang <tangyouling@kylinos.cn>
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>

ei_update_lock is largely vestigal and will probably be removed, but
we're not ready for that just yet.

this fixes some lockdep splats with the new lockdep support for btree
node locks; they're harmless, since we were taking ei_update_lock before
actually locking any btree nodes, but "any btree nodes locked" are now
tracked at the btree_trans level.
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>

Add a pretty printer so the btree reserve cache can be seen in sysfs; as
it pins open_buckets we need it for tracking down open_buckets issues.
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>

another debugging knob
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>

another debugging knob - trigger the journal to do ready journal writes
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>

better stack traces
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>

same as in io_write.c, if we're waiting on the allocator for an
excessive amount of time, print what's going on
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>

Merge tag 'kbuild-fixes-v6.10-4' of git://git.kernel.org/pub/scm/linux/kernel/git/masahiroy/linux-kbuild

Pull Kbuild fixes from Masahiro Yamada:

 - Make scripts/ld-version.sh robust against the latest LLD

 - Fix warnings in rpm-pkg with device tree support

 - Fix warnings in fortify tests with KASAN

* tag 'kbuild-fixes-v6.10-4' of git://git.kernel.org/pub/scm/linux/kernel/git/masahiroy/linux-kbuild:
  fortify: fix warnings in fortify tests with KASAN
  kbuild: rpm-pkg: avoid the warnings with dtb's listed twice
  kbuild: Make ld-version.sh more robust against version string changes

When a software KASAN mode is enabled, the fortify tests emit warnings
on some architectures.

For example, for ARCH=arm, the combination of CONFIG_FORTIFY_SOURCE=y
and CONFIG_KASAN=y produces the following warnings:

    TEST    lib/test_fortify/read_overflow-memchr.log
  warning: unsafe memchr() usage lacked '__read_overflow' warning in lib/test_fortify/read_overflow-memchr.c
    TEST    lib/test_fortify/read_overflow-memchr_inv.log
  warning: unsafe memchr_inv() usage lacked '__read_overflow' symbol in lib/test_fortify/read_overflow-memchr_inv.c
    TEST    lib/test_fortify/read_overflow-memcmp.log
  warning: unsafe memcmp() usage lacked '__read_overflow' warning in lib/test_fortify/read_overflow-memcmp.c
    TEST    lib/test_fortify/read_overflow-memscan.log
  warning: unsafe memscan() usage lacked '__read_overflow' symbol in lib/test_fortify/read_overflow-memscan.c
    TEST    lib/test_fortify/read_overflow2-memcmp.log
  warning: unsafe memcmp() usage lacked '__read_overflow2' warning in lib/test_fortify/read_overflow2-memcmp.c
     [ more and more similar warnings... ]

Commit 9c2d1328 ("kbuild: provide reasonable defaults for tool
coverage") removed KASAN flags from non-kernel objects by default.
It was an intended behavior because lib/test_fortify/*.c are unit
tests that are not linked to the kernel.

As it turns out, some architectures require -fsanitize=kernel-(hw)address
to define __SANITIZE_ADDRESS__ for the fortify tests.

Without __SANITIZE_ADDRESS__ defined, arch/arm/include/asm/string.h
defines __NO_FORTIFY, thus excluding <linux/fortify-string.h>.

This issue does not occur on x86 thanks to commit 4ec4190b
("kasan, x86: don't rename memintrinsics in uninstrumented files"),
but there are still some architectures that define __NO_FORTIFY
in such a situation.

Set KASAN_SANITIZE=y explicitly to the fortify tests.

Fixes: 9c2d1328 ("kbuild: provide reasonable defaults for tool coverage")
Reported-by: Arnd Bergmann <arnd@arndb.de>
Closes: https://lore.kernel.org/all/0e8dee26-41cc-41ae-9493-10cd1a8e3268@app.fastmail.com/Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

After 8d1001f7 (kbuild: rpm-pkg: fix build error with CONFIG_MODULES=n),
the following warning "warning: File listed twice: *.dtb" is appearing for
every dtb file that is included.
The reason is that the commented commit already adds the folder
/lib/modules/%{KERNELRELEASE} in kernel.list file so the folder
/lib/modules/%{KERNELRELEASE}/dtb is no longer necessary, just remove it.

Fixes: 8d1001f7 ("kbuild: rpm-pkg: fix build error with CONFIG_MODULES=n")
Signed-off-by: Jose Ignacio Tornos Martinez <jtornosm@redhat.com>
Reviewed-by: Nathan Chancellor <nathan@kernel.org>
Tested-by: Nathan Chancellor <nathan@kernel.org>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

After [1] in upstream LLVM, ld.lld's version output became slightly
different when the cmake configuration option LLVM_APPEND_VC_REV is
disabled.

Before:

  Debian LLD 19.0.0 (compatible with GNU linkers)

After:

  Debian LLD 19.0.0, compatible with GNU linkers

This results in ld-version.sh failing with

  scripts/ld-version.sh: 18: arithmetic expression: expecting EOF: "10000 * 19 + 100 * 0 + 0,"

because the trailing comma is included in the patch level part of the
expression. While [1] has been partially reverted in [2] to avoid this
breakage (as it impacts the configuration stage and it is present in all
LTS branches), it would be good to make ld-version.sh more robust
against such miniscule changes like this one.

Use POSIX shell parameter expansion [3] to remove the largest suffix
after just numbers and periods, replacing of the current removal of
everything after a hyphen. ld-version.sh continues to work for a number
of distributions (Arch Linux, Debian, and Fedora) and the kernel.org
toolchains and no longer errors on a version of ld.lld with [1].

Fixes: 02aff859 ("kbuild: check the minimum linker version in Kconfig")
Link: https://github.com/llvm/llvm-project/commit/0f9fbbb63cfcd2069441aa2ebef622c9716f8dbb [1]
Link: https://github.com/llvm/llvm-project/commit/649cdfc4b6781a350dfc87d9b2a4b5a4c3395909 [2]
Link: https://pubs.opengroup.org/onlinepubs/9699919799/utilities/V3_chap02.html [3]
Suggested-by: Fangrui Song <maskray@google.com>
Reviewed-by: Fangrui Song <maskray@google.com>
Signed-off-by: Nathan Chancellor <nathan@kernel.org>
Reviewed-by: Nicolas Schier <nicolas@fjasle.eu>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

Pull scheduler fixes from Borislav Petkov:

 - Fix a performance regression when measuring the CPU time of a thread
   (clock_gettime(CLOCK_THREAD_CPUTIME_ID,...)) due to the addition of
   PSI IRQ time accounting in the hotpath

 - Fix a task_struct leak due to missing to decrement the refcount when
   the task is enqueued before the timer which is supposed to do that,
   expires

 - Revert an attempt to expedite detaching of movable tasks, as finding
   those could become very costly. Turns out the original issue wasn't
   even hit by anyone

* tag 'sched_urgent_for_v6.10' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  sched: Move psi_account_irqtime() out of update_rq_clock_task() hotpath
  sched/deadline: Fix task_struct reference leak
  Revert "sched/fair: Make sure to try to detach at least one movable task"

Pull x86 fix from Borislav Petkov:

 - Make sure TF is cleared before calling other functions (BHI
   mitigation in this case) in the SYSENTER compat handler, as
   otherwise it will warn about being in single-step mode

* tag 'x86_urgent_for_v6.10' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  x86/bhi: Avoid warning in #DB handler due to BHI mitigation

Pull i2c fixes from Wolfram Sang:
 "Fixes for the I2C testunit, the Renesas R-Car driver and some
  MAINTAINERS corrections"

* tag 'i2c-for-6.10-rc8' of git://git.kernel.org/pub/scm/linux/kernel/git/wsa/linux:
  i2c: testunit: avoid re-issued work after read message
  i2c: rcar: ensure Gen3+ reset does not disturb local targets
  i2c: mark HostNotify target address as used
  i2c: testunit: correct Kconfig description
  MAINTAINERS: VIRTIO I2C loses a maintainer, gains a reviewer
  MAINTAINERS: delete entries for Thor Thayer
  i2c: rcar: clear NO_RXDMA flag after resetting
  i2c: rcar: bring hardware to known state when probing

Pull smb client fix from Steve French:
 "Small fix, also for stable"

* tag '6.10-rc7-smb3-client-fix' of git://git.samba.org/sfrench/cifs-2.6:
  cifs: fix setting SecurityFlags to true